Highly active antiretroviral therapy (HAART) is rapidly becoming available in the developing world, but the safety and efficacy of different HAART regimens for preventing mother-to-child HIV transmission (MTCT) in a breastfeeding population has not been determined. To test our hypothesis that Trizivir (TZV) may be a simple, safe, and effective strategy to prevent MTCT (PMTCT) in Botswana, 560 pregnant women who intend to breastfeed their infants, and who have CD4 cell counts >200 cells/mm3, will be randomized to receive TZV vs. Lopinavir/Ritonavir (LPV/RTV, or Kaletra)/Combivir (CBV) from 20-34 weeks of pregnancy through 6 months postpartum. An additional group of -140 women with baseline CD4 counts <200 cells/mm3 will receive Nevirapine (NVP)/CBV for treatment, continued indefinitely. All infants will receive single dose (SD) NVP and 1 month of zidovudine (ZDV) prophylaxis. Women will discontinue HAART at 6 months postpartum, or at cessation of breastfeeding if this occurs earlier. If HAART is required for a woman's own health, it will be continued indefinitely with NVP/CBV through the Botswana government's Antiretroviral (ARV) Treatment Program. Women and infants will be followed in the study until infants reach 12 months of age. We believe TZV may perform well when used for PMTCT, and we will determine virologic outcomes by randomization arm to test this hypothesis. We also believe rates of antepartum, intrapartum, and breastfeeding MTCT will be low, and these will be determined among all infants in the randomized study, and among those born to women with CD4 counts <200 cells/mm3 receiving NVP/CBV. Comparison of MTCT rates will be made with a group of infants who received an alternate non-HAART PMTCT strategy in a clinical trial at the same clinic sites. Secondary analyses will complement these primary objectives to provide a complete picture of the risks and benefits of each HAART regimen. Other secondary analysis will address whether HAART started in pregnancy leads to infant prematurity and low birth weight;whether LPV/RTV concentrations are low in pregnancy;and whether low drug concentrations in plasma and breast milk are associated with viral load and resistance mutations in these compartments, and with MTCT. The Botswana government currently offers HAART to all citizens with CD4 cell counts <200 cells/mm3, and is considering offering TZV-based HAART to HIV-infected pregnant women with higher CD4 cell counts for PMTCT. The Botswana government has given explicit support for this protocol to occur in four existing Botswana-Harvard AIDS Institute Partnership (BHP) locations to help guide future policy decisions.